what is Atlassian Crowd

[sn]

hi all,

We have our Confluence user directory configured to JIRA Server and under the Type , it says Atlassian Crowd.

I read some documents but looks like Atlassian Crowd is a software that needs to be installed and there is cost involved. I believe we do not have Crowd.

why is this listed as Atlassian Crowd and how is this directory type different than the Atlassian Crowd ?

thanks !

[Nic Brough -Adaptavist-]

Crowd is an identity system.  There's two versions of it though.  One is a Server which does what you'd expect - manage users and provide a directory for other systems to use.

The other is "embedded crowd" - this is (sort of) a cut down crowd server which provides the "internal" user directory in Jira, Confluence, Bitbucket etc.  It's less functional than server, and has different localised interfaces, so you use it inside the application it is embedded in, rather than as a separate service.

What you have done is told Confluence to use Jira as its directory, rather than its own separate list in its own embedded crowd.  So it's effectively using the Jira embedded Crowd instead of its own.

[Serge Calderara]

@Nic Brough -Adaptavist- thanks for this explaination.

I am exactly in this case using Jira 9.12.2 in on side and confluence 8.9 on the other side where confluence diractory is using Jira as its directory

We have noticed a wierd situation when using it.

In our LDAP we are using Groups and SubGroups, when we sync jira directory, Sub-groups are properly collected and created in Jira.

In an other hand, when we run the confuence Sync, sub groups are not created in confluence side while the enable sub-group is set to true 

Any idea why subgroups goes to Jira but not to confluence ?

regards

[Nic Brough -Adaptavist-]

Niether Confluence nor Jira used to recognise sub-groups, they just flattened them out during the sync (so if you had Group B as a sub-group of A, and a user in only group B, Jira would import Group A and Group B, and add the user to both groups and treat the memberships as totally independent)

I don't think Confluence has ever had any more recognition of sub-groups added to it.

[Serge Calderara]

hello @Nic Brough -Adaptavist- 

PLease not we are using Jira and Confluence DC

What I mean is the following. We have an LPAP with group and sub groups.

Exemple

GroupA

       Group A1 ( sub group) : user John

       Group A2 ( sub group): user Bob

In Jira the Directory sync to our LDAP has Enabled nested group=true and when sync start, we can see in Jira  John and Bob which are respectivly member of Group A1 and Group A2

In confluence side , which use Jira crowd diretory Type and use Jira directory above, then when synking, we can see both user but we cannot see than they belongs to Group A1 and Group A2  even with Enabled nested group = true

As in confluence we are using those Group A1 and Group A2 to apply permission in spaces for theire memebers, 

Any idea ?

regarsd

[Nic Brough -Adaptavist-]

If you are using crowd, you are using DC (or server, which is no longer supported, but effectively the same as DC).

As I said, it is just flattening them as it reads them from your LDAP, it doesn't do sub-groups.

[Serge Calderara]

hello @Nic Brough -Adaptavist- ,

I think I miss explain or you miss understand.

We are not using crowd server in confluence directory for user management, we are using an embeded version of crown in a way that is is using the Jira active directory sync configuration.

What I mean is that at our LDAP, group A and Group B are sub groups of GoupeParent

When jira directory sync, which is directly connected to our LDAP, then Group A and Group B are going down and flatted as you said which is correct. So we can see both groups in user management

When confluence sync, I am suppose to see also those Group A and Group B flatted and visible BUT in fact they are not going down to confluence, only GroupParent is seen

Hope you understand better what I mean

Why in confluence GroupA and B are not created as flatten group and visible in same way as in Jira ?

regards

[Nic Brough -Adaptavist-]

You had not said you were using internal Crowd, so yes, there was not much I could tell you.

It is the same answer though - one system flattens them properly, the other does not because it does not recognise sub-groups.